I thought what Michael Whapples wrote was excellent and I agree that
different braille users can have good reasons for making different choices
and also that the use of a braille display rather than paper braille can
affect these choices. The various possible choices don't of course eliminate
the need to support standard braille systems. The only thing I would add to
what Michael wrote about braille rules is a reminder that Nemeth braille
math is much simpler and more reflective of print math than is UEB math.
Just to underline what both Michael and Neil Soiffer posted, AsciiMath only
represents a subset of MathML so direct translation from AsciiMath to
braille math doesn't address the need to translate MathML to braille as is
required by ePUB as well as other major file formats.
I have a lot of experience developing accurate backtranslators from both
Nemeth and UEB braille math to print math. In my experience most of the
difficulty is unrelated to the particular representation of the output print
math.
Backtranslation or interpretation of braille requires what is known in
computer science as lexical analysis and parsing which is not the same as
table lookup. I use the ANTLR 4 parser generator to automatically generate
both a lexer and a parser from my grammars for a braille math system. The
parser outputs a parse tree which is then backtranslated using a custom
procedure. I use Java for the latter but ANTLR 4 also supports a number of
other programming languages.
Best, Susan Jolly
For a description of the software, to download it and links to
project pages go to http://liblouis.org
Donate: http://liblouis.org/sponsoring
